Who are you?

Introduce “grammar of graphics” structure
Practice!
- samples with built-in data
- preparing data
- samples with (more interesting?) data
Send you away with a list of resouces (including these slides)

Code structure:

Grammar of graphics

ggplot(data = diamonds, aes(x = carat, y = price, color = clarity)) +
  geom_point() +
  facet_grid(color ~ cut)

Let’s try it!

Download scripts from the following site: https://is.gd/IYoXwA
Open RStudio.
In RStudio, open “install_packages.R”. Highlight the text and click “Run”.
Still in RStudio, open “workshop_script.R”. We’ll work from this for the rest of the presentation.
Load packages with the “library” commands at the top of the script.

Okay, now let’s try it!

First, we need data.

Let’s use the built-in R dataset, “diamonds”.

carat	cut	color	clarity	depth	table	price	x	y	z
0.23	Ideal	E	SI2	61.5	55	326	3.95	3.98	2.43
0.21	Premium	E	SI1	59.8	61	326	3.89	3.84	2.31
0.23	Good	E	VS1	56.9	65	327	4.05	4.07	2.31
0.29	Premium	I	VS2	62.4	58	334	4.20	4.23	2.63
0.31	Good	J	SI2	63.3	58	335	4.34	4.35	2.75
0.24	Very Good	J	VVS2	62.8	57	336	3.94	3.96	2.48

Data

ggplot(data = diamonds)

Aesthetics: aes()

ggplot(data = diamonds,
       aes(x = carat, y = price))

Geometric object: geom_

ggplot(data = diamonds, 
       aes(x = carat, y = price)) +
  geom_point()

Add more aesthetics…

ggplot(data = diamonds, 
       aes(x = carat, y = price, color = clarity)) +
  geom_point()

Why didn’t we do this?

ggplot(data = diamonds, 
       aes(x = carat, y = price), color = clarity) +
  geom_point()

We could get this instead:

ggplot(data = diamonds, 
       aes(x = carat, y = price)) +
  geom_point(color = "magenta")

Our original plot

(with a small difference:)

ggplot(data = diamonds, 
       aes(x = carat, y = price)) +
  geom_point(aes(color = clarity))

Add more geoms:

ggplot(data = diamonds, 
       aes(x = carat, y = price)) +
  geom_point(aes(color = clarity)) +
  geom_smooth()

Change geom appearance:

ggplot(data = diamonds, 
       aes(x = carat, y = price)) +
  geom_point(aes(color = clarity)) +
  geom_smooth(color = "black", size = 0.8, linetype = 2)

Make it look nicer:

ggplot(data = diamonds, 
       aes(x = carat, y = price)) +
  geom_point(aes(color = clarity)) +
  geom_smooth(color = "black", size = 0.8, linetype = 2) +
  theme_few()

Try facets!

ggplot(data = diamonds, 
       aes(x = carat, y = price)) +
  geom_point(aes(color = clarity)) +
  geom_smooth(color = "black", size = 0.8, linetype = 2) +
  facet_wrap(~cut)

Two-dimensional facets:

ggplot(data = diamonds, aes(x = carat, y = price, color = clarity)) +
  geom_point(size = 0.5) +
  facet_grid(color~cut)

Try out color scales

ColorBrewer is useful and popular:

ggplot(data = diamonds, aes(x = carat, y = price)) +
  geom_point(aes(color = clarity)) +
  theme_few() + 
  scale_color_brewer(type = "qual", palette = "Set2")

Different plot types?

ggplot(data = diamonds, 
       aes(x = clarity, y = price)) +
  geom_violin()

Try an extension:

ggplot(data = diamonds, 
       aes(x = price, y = cut)) +
  geom_joy()

Choose colors: scale_color_…

ggplot(data = diamonds, 
       aes(x = price, y = cut, color = cut, fill = cut)) +
  geom_joy(alpha = 0.8, scale = 5) +
  scale_fill_viridis(option = "A", discrete = TRUE) +
  scale_color_viridis(option = "A", discrete = TRUE) + 
  theme_few()

Another color scale:

ggplot(data = diamonds, 
       aes(x = price, y = cut, fill = cut)) +
  geom_joy(alpha = 1, scale = 5) +
  scale_fill_manual(values = wes_palette("Darjeeling")) +
  theme_few()

Sometimes we need to transform data.

kable(head(iris))

Sepal.Length	Sepal.Width	Petal.Length	Petal.Width	Species
5.1	3.5	1.4	0.2	setosa
4.9	3.0	1.4	0.2	setosa
4.7	3.2	1.3	0.2	setosa
4.6	3.1	1.5	0.2	setosa
5.0	3.6	1.4	0.2	setosa
5.4	3.9	1.7	0.4	setosa

Wide or long?

iris_long <- melt(iris, id.vars = ("Species"))
kable(head(iris_long))

Species	variable	value
setosa	Sepal.Length	5.1
setosa	Sepal.Length	4.9
setosa	Sepal.Length	4.7
setosa	Sepal.Length	4.6
setosa	Sepal.Length	5.0
setosa	Sepal.Length	5.4

Plot the data:

ggplot(iris_long, aes(x = Species, y = value, fill = variable)) +
  geom_bar(stat = 'identity', width = 1) +
  theme_bw()

Change coordinates

ggplot(iris_long, 
       aes(x = Species, y = value, color = variable, fill = variable)) +
  geom_bar(stat = 'identity', width = 1) +
  coord_polar(theta = 'x') +
  theme_bw()

A more interesting dataset:

kable(head(gapminder))

country	continent	year	lifeExp	pop	gdpPercap
Afghanistan	Asia	1952	28.801	8425333	779.4453
Afghanistan	Asia	1957	30.332	9240934	820.8530
Afghanistan	Asia	1962	31.997	10267083	853.1007
Afghanistan	Asia	1967	34.020	11537966	836.1971
Afghanistan	Asia	1972	36.088	13079460	739.9811
Afghanistan	Asia	1977	38.438	14880372	786.1134

A scatter plot:

ggplot(gapminder, aes(x = gdpPercap, y = lifeExp)) +
  geom_point()

Who’s the outlier?

gapminder %>% filter(gdpPercap > 60000)

## # A tibble: 5 x 6
##   country continent  year lifeExp    pop gdpPercap
##    <fctr>    <fctr> <int>   <dbl>  <int>     <dbl>
## 1  Kuwait      Asia  1952  55.565 160000 108382.35
## 2  Kuwait      Asia  1957  58.033 212846 113523.13
## 3  Kuwait      Asia  1962  60.470 358266  95458.11
## 4  Kuwait      Asia  1967  64.624 575003  80894.88
## 5  Kuwait      Asia  1972  67.712 841934 109347.87

Deal with overplotting:

ggplot(gapminder, aes(x = gdpPercap, y = lifeExp)) +
  geom_hex()

Another way to avoid overplotting:

ggplot(gapminder, aes(x = gdpPercap, y = lifeExp)) +
  geom_density2d(aes(color = ..level..), bins = 20) +
  scale_color_viridis()

A little more complicated:

p <- ggplot(gapminder, aes(x = gdpPercap, y = lifeExp)) +
  geom_point(aes(color = continent, size = pop), alpha = 0.8) +
  scale_x_continuous(trans = 'log') +
  facet_wrap(~year) +
  scale_color_brewer(type = "Qual", palette = "Accent") +
  theme_hc(bgc = 'darkunica') +
  theme(text = element_text(size = 9))

A little more complicated:

Let’s make a map!

Prepare data:

country_df <- map_data('world') %>%
  rename("country" = "region") 
country_df$country[country_df$country == "USA"] <- "United States"

#Take the mean across all years for each country:
gapminder_means <- gapminder %>% 
  group_by(country, continent) %>%
  summarise(lifeExp = mean(lifeExp),
            pop = mean(pop),
            gdpPercap = mean(gdpPercap))

plot_dat <- left_join(gapminder_means, country_df, by = "country")

Make the map:

ggplot(plot_dat) +
  geom_polygon(aes(x = long, y = lat, fill = lifeExp, group = group)) +
  scale_fill_viridis(option = "A") + 
  coord_quickmap() +
  theme_few()

More plot ideas?

What questions could we ask with this data?
How could we visually answer those questions?

Where to learn more:

Stack Overflow/Google
Books:
- R for Data Science: http://r4ds.had.co.nz/program-intro.html
- Ggplot2: Elegant graphics for data analysis: https://www.amazon.com/ggplot2-Elegant-Graphics-Data-Analysis/dp/0387981403 (may also be available for free at github).
- Data visualization for social science: http://socviz.co/

Where to learn more:

Papers:
- Wickham grammar of graphics: http://byrneslab.net/classes/biol607/readings/wickham_layered-grammar.pdf
- Wickham tidy data: https://www.jstatsoft.org/article/view/v059i10
Tutorials, cheatsheets, websites:

Where to learn more:

Twitter:
- @hadleywickham, @ClausWilke, @JennyBryan, @RLadiesGlobal, @rstudiotips
This presentation! https://adrienne-marshall.github.io/ggplot2_workshop/index.html

Questions?

Thanks!

Adrienne Marshall mars7850@vandals.uidaho.edu

Beautiful Graphics with ggplot2

Adrienne Marshall

9/12/2017

Who are you?

Who am I?

What is ggplot2?

Why ggplot2?

London population over time:

Single-nucleotide polymorphisms:

Measles occurrence in 50 states

Europe’s population

Global remittances

Pretty plant drawing

London cycle hires

Relationship status by age

Goals for today

Goals for today

Code structure:

Let’s try it!

Okay, now let’s try it!

Data

Aesthetics: aes()

Geometric object: geom_

Add more aesthetics…

Why didn’t we do this?

We could get this instead:

Our original plot

Add more geoms:

Change geom appearance:

Make it look nicer:

Try facets!

Two-dimensional facets:

Try out color scales

Different plot types?

Try an extension:

Choose colors: scale_color_…

Another color scale:

Sometimes we need to transform data.

Wide or long?

Plot the data:

Change coordinates

A more interesting dataset:

A scatter plot:

Who’s the outlier?

Deal with overplotting:

Another way to avoid overplotting:

A little more complicated:

A little more complicated:

Let’s make a map!

Make the map:

More plot ideas?

Where to learn more:

Where to learn more:

Where to learn more:

Questions?